Improving Tropical Cyclone Forecasts from Formation to Maturity Using Ensemble-Based Data Assimilation
Open Access
- Author:
- Hartman, Christopher
- Graduate Program:
- Meteorology and Atmospheric Science
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- February 08, 2024
- Committee Members:
- Paul Markowski, Program Head/Chair
Steven Greybush, Major Field Member
Eugene Clothiaux, Major Field Member
Manzhu Yu, Outside Unit & Field Member
Xingchao Chen, Chair & Dissertation Advisor - Keywords:
- Data Assimilation
Ensemble Kalman Filter
Meteorology
Numerical Weather Prediction
Remote Sensing
Satellites
Tropical Cyclones
Tropical Meteorology - Abstract:
- The studies comprising this dissertation use a state-of-the-art ensemble-based data assimilation (DA) system developed at The Pennsylvania State University to improve forecasts of tropical cyclones (TCs) during two of the least predictable stages of their lifecycle: formation (i.e., tropical cyclogenesis; hereafter TCG) and rapid intensification (RI). These improvements are realized by assimilating infrared (IR) brightness temperatures (BTs) observed by geostationary satellites under both clear and cloudy conditions. The all-sky IR BTs assimilated by the DA system help to constrain the initial moisture estimates within the core of the developing system in analyses via the strong ensemble correlations that exist between moisture content and simulated BTs. It is shown that forecasts initialized from these analyses exhibit a more realistic convective evolution, which translates to improved prediction of TCG and RI. For the case of TCG, the assimilation of upper-tropospheric water vapor channel BTs observed by the Meteosat-10 Spinning Enhanced Visible and InfraRed Imager (SEVIRI) instrument improves the timing of TCG in forecasts of Hurricane Irma (2017). In an experiment that withheld the BTs, TCG was premature by at least 24 hours due to an overestimation of the spatial coverage of deep convection within the African Easterly Wave (AEW) that Irma formed from. Spurious convection led to stronger low-level convergence and the earlier spin-up of a low-level meso-β-scale (i.e., 20 – 200 km) vortex. This was ameliorated by assimilating all-sky IR BTs. Furthermore, the substantial impact of initial moisture uncertainty within the incipient disturbance is revealed by initializing ensemble forecasts with only the initial moisture perturbations retained. Relative to an ensemble with initial perturbations to all variables, at least half of the intensity forecast uncertainty is attributed to initial moisture uncertainty within the AEW. These results show the importance of targeting the incipient disturbance with high spatio-temporal water vapor observations for ingestion into DA systems. For the case of RI, the assimilation of upper-tropospheric water vapor channel BTs observed by the GOES-16 Advanced Baseline Imager (ABI) led to significant improvements in the intensity forecasts of Hurricane Dorian (2019) at lead times of 48 hours and longer. These improvements are shown to be a result of better analyzed cloud fields as well as more intense initial primary and secondary circulations. Despite these improvements, the vortex exhibited an unrealistically broad structure that was fine-tuned by the additional assimilation of tail Doppler radar (TDR) radial velocities collected by NOAA P-3 aircraft. The simultaneous assimilation of all-sky IR BTs and radar observations therefore resulted in realistic forecasts of the track, structure, and RI of Dorian. These results underscore the potential of TDR observations to complement the benefits gained by assimilating all-sky IR BTs.